Cognition and State Modeling

R3LIVE

R3LIVE is a tightly-coupled LiDAR-inertial-visual state estimation and mapping tool that reconstructs robust RGB-colored 3D maps in real time.

Tool Introduction

Core parameters, trigger timing, and visual before/after demo references.

Short Explanation

Run synchronized LiDAR, IMU, and camera streams and R3LIVE estimates state while producing a real-time RGB-colored 3D map.

InputLiDAR + IMU + camera stream

OutputState estimate, RGB point map, textured reconstruction

Trigger TimingTriggered when the ROS launch file receives synchronized sensor streams.

RuntimeROS / C++ / FAST-LIO + visual-inertial mapping

BeforeLiDAR + IMU + camera stream

Prepare the scene, image, video, sensor stream, prompt, or configuration expected by the original project.

AfterState estimate, RGB point map, textured reconstruction

Read the produced visualization, prediction, map, trajectory, mask, grasp pose, or other documented artifact.

Preset Example

A quick-run style example for the documentation page.

Inputtools/r3live/datasets/YOUR_DOWNLOADED.bag

PromptRun r3live_bag.launch and visualize the colored map

ExpectedA state trajectory, RGB-colored point map, and optional mesh reconstruction.

Parameters And Output

Readable controls and the meaning of each returned artifact.

Parameter Explanation

launch_fileselectr3live_bag.launch

ROS launch entry for the dataset or live sensor setup.

configpath

Sensor calibration, camera model, LiDAR topic, and mapping parameters.

rosbagfile

Recorded LiDAR-inertial-visual stream to replay.

mesh_reconstructiontogglefalse

Runs the additional reconstruction utility after mapping.

Output Explanation

state_estimate

Estimated pose, velocity, and sensor state used for localization.

rgb_point_map

LiDAR map colored by camera information for readable scene reconstruction.

mesh

Optional reconstructed surface generated from the map output.

How To Use

Official resources, deployment steps, academic context, citation, and source-reported benchmark numbers.

Resources

GitHubhttps://github.com/hku-mars/r3live Code Downloadhttps://github.com/hku-mars/r3live/archive/refs/heads/master.zip Paper PDFhttps://github.com/hku-mars/r3live/blob/master/papers/R3LIVE%20--%20A%20Robust%2C%20Real-time%2C%20RGB-colored%2C%20LiDAR-Inertial-Visual%20tightly-coupled%20stateEstimation%20and%20mapping%20package.pdf R3LIVE Datasethttps://github.com/ziv-lin/r3live_dataset Mesh Reconstructionhttps://github.com/hku-mars/r3live#mesh-reconstruction Videohttps://youtu.be/j5fT8NE5fdg

Deployment Notes

Install ROS and the R3LIVE dependencies listed in the official repository.
Build the catkin workspace and verify the Livox/FAST-LIO related packages are available.
Download the official datasets or prepare calibrated live sensor topics.
Launch R3LIVE, replay the rosbag, and save RGB map or mesh outputs for inspection.

Relative Path Example

# Relative-path local entry for the R3LIVE tool folder
cd tools/r3live/catkin_ws
catkin_make
source devel/setup.bash

roslaunch r3live r3live_bag.launch
rosbag play tools/r3live/datasets/YOUR_DOWNLOADED.bag

# Mesh reconstruction utility:
roslaunch r3live r3live_reconstruct_mesh.launch

# Suggested repository layout:
# tools/r3live/README.md
# tools/r3live/r3live/launch/r3live_bag.launch
# tools/r3live/config/
# tools/r3live/datasets/

# This page documents the path. The static page does not execute R3LIVE.

Expected Result Shape

{
  "tool": "r3live",
  "status": "ok",
  "trajectory": [
    {
      "label": "RGB-colored LIV mapping",
      "score": 0.87,
      "output": "State estimate, RGB point map, textured reconstruction"
    }
  ],
  "timing": {
    "runtime": "PC VIO per-frame cost is 7.01 ms at 320x256 / 0.10 m point resolution, 11.53 ms at 640x512 / 0.10 m, and 13.63 ms at 1280x1024 / 0.10 m; LIO per-frame cost is 18.40 ms.",
    "device": "documented in source benchmark when available"
  },
  "artifacts": {
    "visualization": "tools/r3live/runs/visualization.png",
    "raw_predictions": "tools/r3live/runs/predictions.json"
  }
}

Paper figure

Academic Info

Paper identity and contribution summary.

TitleR3LIVE: A Robust, Real-time, RGB-colored, LiDAR-Inertial-Visual tightly-coupled state Estimation and mapping package

AuthorsJiarong Lin, Chunran Zheng, Wei Xu, Fu Zhang

VenueICRA 2022

ContributionCombines LiDAR-inertial odometry with visual-inertial color rendering to produce robust state estimates and dense RGB-colored maps in challenging environments.

Citation

@misc{r3live2022,
  title={R3LIVE: A Robust, Real-time, RGB-colored, LiDAR-Inertial-Visual tightly-coupled state Estimation and mapping package},
  author={Jiarong Lin and Chunran Zheng and Wei Xu and Fu Zhang},
  year={2022},
  note={ICRA 2022},
  url={https://github.com/hku-mars/r3live/blob/master/papers/R3LIVE%20--%20A%20Robust%2C%20Real-time%2C%20RGB-colored%2C%20LiDAR-Inertial-Visual%20tightly-coupled%20stateEstimation%20and%20mapping%20package.pdf}
}

Benchmark

Only compact, source-reported numbers are shown here.

Dataset	Metric	Value	Runtime	Source
HKUST campus loops	Loop drift	0.093 m, 0.154 m, 0.164 m, 0.102 m over 1.19-1.52 km trajectories	Real-time mapping pipeline	ICRA 2022 paper
Runtime table	Per-frame time	VIO 7.01 ms at 320x256 / 0.10 m, LIO 18.40 ms	PC evaluation	R3LIVE paper

Artifacts

R3LIVE paper, drift table, RPE table, runtime table, ROS launch files, dataset bags, RGB map outputs, and mesh reconstruction utilities.

Demo Images

Visual references from the original tool. Click any image to inspect the original size.